SRU optimization for predictive maintenance on RISC-V

We are pleased to present our new contribution to embedded systems research: “Optimizing SRU Models for Predictive Maintenance on Embedded Systems (RISC-V).”

The challenge: How to efficiently deploy machine learning models for predictive maintenance on embedded systems with strict power and resource constraints?

This challenge is crucial for industries using critical systems such as aeronautics, automotive or energy.

Our solution: We have optimized SRU (Simple Recurrent Unit) models for execution on RISC-V processors, developing specific techniques to reduce memory footprint, improve real-time performance, and minimize power consumption.

This work shows how to reconcile predictive model performance and hardware requirements on embedded platforms.

Possible applications :
– Condition monitoring for industrial systems
– IoT predictive maintenance
– Energy optimization for embedded devices

Want to find out more?
Read the full article, or take a look at the two-page summary.